Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+21 moreClaude CodeCursorCodex CLIMuseDots
Manus
Grok
Grok Bot+18 agentsVerification-Before-Done
It treats verification as a first-class artifact, not an afterthought.
$8.99
Verification-Before-Done
Example session with this skill installed
I fixed a bug where our Node/Express app sent password reset emails twice. The job was queued in both the controller and a Mongoose post-save hook, and I removed the hook. Is it actually done?
- Read your context and instructions
- Compiled the verification-before-done
Verification plan for the duplicate password reset email fix. It is not done until the checks below produce evidence.
Definition of "working"
- One reset request produces exactly one job in the queue and one email sent
- Other flows that relied on the post-save hook still behave as before
- The reset link in the email still works end to end
Checks (evidence over inference)
| # | Check | Evidence to capture |
|---|---|---|
| 1 | Trigger one reset in staging | Queue shows 1 job for that user ID |
| 2 | Inspect mail provider log | 1 message to the test address |
| 3 | Search the codebase for other callers of the hook's function | List of call sites, or "none" |
| 4 | Save a user through a non-reset path (profile update) | No reset job queued |
| 5 | Click the reset link, set a new password | Login with the new password succeeds |
| 6 | Add a test asserting the enqueue is called once | Test fails on the old code, passes on the new |
"The hook is gone so it must be fixed" is inference. Checks 1, 2 and 6 turn it into evidence.
Evidence trail
Record per check: date, environment, command or action, observed result, link to log or screenshot. Attach it to the PR.
Next steps
- Run checks 1 to 5 in staging and fill in the trail
- Write the regression test from check 6
- Mark done only when every row has evidence
Connects securely to your tools. The creator never sees your data.
About this skill
'Should work now' is the most expensive sentence in software — it means the change was reasoned about, not observed, and the next person to touch this code will pay for that assumption. This skill builds the verification protocol for your specific change: the exact commands to run, the outputs to observe, the edge cases to test manually, and the signals that confirm the fix didn't introduce a regression. It treats verification as a first-class artifact, not an afterthought. Use it before you close any ticket — give it your change, your environment, and what you've already tested; it returns the complete verification checklist that lets you say 'done' and mean it.
- Cannot execute live code to physically verify results.
- Relies on user providing accurate environment and stack details.
- Does not replace automated CI/CD pipelines.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 4 months ago
Permissions
No special permissions declared or detected